INCIDENCE OF LANSLIDE HAZARD: ITS MAPPING USING REMOTE SENSING, GIS AND LOGISTIC REGRESSION MODEL IN EYINOKE HILLY AREA OF OKEIGBO, SOUTHWESTERN NIGERIA

Occurrences of landslides are of considerable constraints to social and environmental developments in hilly/mountainous areas of the world. Okeigbo, a settlement in southwest of Nigeria, situated on the hilly terrain was recently affected by a major landslide in Eyinoke area and destroyed many infrastructures and livelihoods. This study involved landslide susceptibility assessment and its quantification of hazard and risk related in the area using remote sensing imageries, geographic information system (GIS) and logistic regression model. Landsat 7 imagery of 2006 was analysed using geospatial analytical software such as Arc GIS 10.0, Erdas Imagine and Global Mapper 13.0 for zonation and statistical analyses respectively. The logistic regression model was used to find the best fitting function to establish the relation between dependent variables (presence or absence of landslides) and a set of independent variables such as lithology, slope angle, geomorphology, land use, drainage density, lineament density, weathering, proximity to road. Overlay analysis was carried out by evaluating the layers obtained according to their accepted coefficient in final model. Based on the results, the study area was classified into five classes of relative landslide susceptibility which include viz: low, moderate, high and very high. Majority of the areas on the hill, east of Okeigbo are under the threat of landslide because the vegetation has being largely removed thereby exposing the soil to the heavy rainfalls. Other areas prone to landslide in the study area include viz: cultivated slope, road and residential area. It was also observed that the recent landslide in Okeigbo results from combination of different factors which the logistic regression model predicted to include slope, proximity to road, river and residential. The slopes in the area were steep and close to drainage structure which affected their stability. Similarly, rivers and small streams in the area may have adversely affected stability by eroding the slopes hence contributing to the recent landslide. Verification of results of the analysis using data collected from present landslide location and its comparison between the predicted result and actual distribution of landslides showed a high consistency.